The mammalian auditory brainstem is a complex signal-processing device composed of multiple interconnected neural circuits. Signal transformations in these circuits underlie the perceptual attributes of sound, many of which depend on us having two spatially-separated ears. Sound-source localization and vocal communication in noisy environments both rely heavily on neural sensitivity to micro-second differences in the timing of sounds at the two ears. This inter-aural time sensitivity first emerges in neurons of the medial superior olive (MSO), in the ventral brainstem. Each MSO neuron receives spectrally-tuned and temporally precise inputs from both ears, and is itself “tuned” to a preferred inter-aural time delay; its “best delay”. How tuning for inter-aural delay is achieved is fiercely debated; hypotheses abound, while data are scarce. Using systems-identification techniques, we test the hypothesis that the required “internal time delays” result from mismatches in frequency tuning between the two ears’ inputs to MSO neurons.

We recorded spike times in response to inter-aurally uncorrelated broadband Gaussian noise from a large population of electrically-isolated MSO axons in the anesthetized chinchilla. From a spike-triggered reverse correlation analysis, we estimated the impulse response of both ears’ independent inputs to this binaural system. We found significant frequency-tuning mismatches between ipsi- and contra-lateral inputs to many MSO neurons, resulting in frequency-dependent internal neural-circuit delays. For each single neuron, a linear model based on time-domain cross correlation of its two monaural filters accurately predicted many of that neuron’s response properties to other (independent) binaural sounds. At the population level, the distribution of frequency-tuning mismatches predicted the observed distribution of best delays. However, the model failed to account for some aspects of the population data; perhaps indicating an additional source of internal delay for some neurons.

We conclude that frequency-tuning disparities between the two ears’ inputs to MSO neurons are common, and that they can account for the main binaural properties of this circuit. The translational implications of these findings become clear when considering that people with hearing loss often have impaired cochlear frequency tuning, which can alter inter-aural phase delays at the input to binaural circuits. Hearing-impaired people have specific perceptual deficits related to binaural hearing (e.g., difficulties following conversations in complex acoustic environments, such as restaurants), for which current auditory prostheses do little to help. Application of our detailed binaural-circuit analyses to the design of future implantable assistive-device technology has potential to restore functional binaural advantages for hearing-impaired listeners.

BME Faculty Host: Pedro Irazoqui~

***Coffee and juice will be provided at West Lafayette***